I -IVS*

I -IVS*

I

-pre-mANA

-mANA

-E1

7

Fig 3A. Chasing of heat inactivated (~) prp18 spliceosomes

The gradient prome of a 50ml splicing reaction conducted with inactivated prp18 extracts.

The reaction mixture was sedimented for 100 min. and analyzed as described in Materials and Methods. Analysis of the labeled RNA in an aliquot of the gradient fractions (data not shown) revealed a peak of lariat intermediate at fraction 12, corresponding to the peak of 32p labeled RNA in the 40S region of the gradient.

Fig 3B. Complementation of prp18 (D) spliceosomes. Fraction 12 from the gradient de- scribed in panel A, was used as the prp18 heat inactivated spliceosome. The chasing of the intermediates in the spliceosome was done as described in Materials and Methods. No complementing extract was added in lane 7, lane 1 was incubated with inactivated prpll , lane 2 with inactivated prp5, lane 3 with fraction 40P3 from wild type extract, lane 4 with micrococcal nuclease-treated 40P3, lane 5 with 40W fraction of wild type extract, lane 8 was incubated with inactivated prpll extract in the absence of ATP in the chasing mix.

The products of the chasing reaction, lariat intron (IVS*) and the spliced exons (mRNA), are indicated in the figure.

~

:~J a..

()

A ji

9 11 13 15 17 19

Fraction Number

B z

<] ~

z ....

..-<](")(")~::Q)

,_ 'g.

Q_ Q_ I :t:

e-

ooWo::J Q.Q."O;t"O;t()"Otal

.. .

1VS*E2-~~;._ .. ____ f'll 1VS*-____ ..._ __ ·

~

' _,..._

-

__

... . .

.. ' pre-mANA- mANA-·-... -' " ..,. ~~·

25 Top

E1- .1 2 3 4 5 6 7

a..

t{

I <] ... ...

~ - -

Vl ~ 8

Fig 4. In vivo complementation analysis with clones and subclones of PRP 18 .

Restriction map ofpUV18 and complementation phenotype for the subclones with portions of the genomic DNA fragment in pUV18. Only restriction sites relevant to the isolation of subclones and the generation of hybridization probes are indicated. + indicates ability to suppress the temperature sensitive growth defect of prp18, and- incidates inability to con- fer temperature insensitivity. The wavy line represents YCp50 sequences and the hatched box the complementing region. Restriction enzymes: B, Bam Ill; C, Cia I; E, EcoRI; H, Hindlll; N, Nru I, S, Sal I and X, Xho I.

Plasmid

pUV18-17 pUV18-19 pUV18·3 pUV18·2 pUV18·4 pUV18·20 pUV18 pUV18-15 pUV18·15.8 pUV18·15.7 pUV18-15.2 pUV18·15.1

E B N

vcpsovJvvvJ , r s

< I ... ... I

c

X H ... I '

·~

Complementation

Scale 1kb W\1\A Scale 0.5kb

+ + + + + +

Fig 5A. Complementation of lariat intermediate accumulation phenotype of prp18

Total RNA (lOmg in each case) prepared from cultures at 230C and 370C was elec- trophoresed through a 1.2% formaldehyde agarose gel, blotted and crosslinked to Gene- Screen. The blot was then probed with a labeled fragment of the actin intron. Lanes 1 and 2 were loaded with RNA from a prp18 strain transformed with the vector YCp50, lanes 3 and 4 with RNA fromprp18 transformed with the complementing subclone pUV18, lanes 5 and 6 prpl8 transformed with complementing subclone pUV18-15 and lanes 7 and 8 with RNA from the untransformed prp18 strain. The position of the actin lariat intermedi- ate is indicated.

Fig 5B. Suppression of lariat intermediate phenotype of RP51A

Reverse transcription was done to map the 5' end of RP51A transcripts. 20mg of total RNA was hybridized to a RP51A oligonucleotide complementary to sequences 70nts downstream from the 5' splice site. The extension products are separated on 10% 8M urea aery lamide gels. RNA from untransformed wild type and prp 18 strains, the prp 18 strain transformed with the complementing plasmid pUV18, prp18 strain transformed with the subclone pUV18-15, prpl8 strain transformed with a complementing multicopy plasmid

were used. The last lane contains kinased oligonucleotide markers of sizes 70, 55 and 35 nts. The strong extension stop seen in ts 503 just above the 70 nts marker corresponds to a reverse transcription stop at the 5' end of the accumulated lariat intermediate in the mutant strain.

1\.)

• :] ts503YCp50 )>

w w

•

-...! 1\.)

:] ts503pUV18

w w -...! 1\.)

~] ts503pUV18-15

w w ^-...! 1\.) w

: J ts503

w -...!

I <

(/)

• m

f\)

SS330 CJ ts503pUV18 ts503pUV18-15

-------

ts503pUV18-15(2JL) ts503 • • '

~ I I I

w

U1 ""-J 01 U1 0

Fig 6A. Integration of the cloned DNA to the homologous choromosomal PRP 181ocus Cartoon of the expected integration event. A 5 kb complementing Eco RI fragment (see Fig 4) subcloned in YIP5 was linearized at the unique Xho I site and transformed in the wild type strain SS330. The open box represents the chromosomal locus and the cloned yeast DNA is represented by the hatched box. Plasmid sequences are represented by thin solid lines and markers in the plasmid as labeled boxes (Ap represents ampicillin resistance, Tc represents tetracycline resistance and URA3 is the yeast marker gene). Restriction sites are represented by the same notations as in Fig. 4.

Fig 6B. Southern blot confirming integration event at the cloned genomic locus. DNA from wild type, prp18 strain and from two integrants SS330 Ia and Ib was digested with Sal I and run on a 1.0% agarose gel, blotted and probed with the 4.5 kb Bamlll to EcoRI (see Fig. 4) complementing fragment.

A

s

X N

t:::/.

160

s ~ ~~~===±=====±======~

X

X N E

s

•

E X N

s

--

~~

oC':J66

("') CX) ("') ("') ("') ... ^("') ("')

8

(/)e-CI)(/)

(/)Q..CI)(/)

-

E

s

Fig 7. Mapping the PRP 18 transcript.

A restriction map of the complementing fragment containing the PRP 18 gene is represented by the open box, and the wavy lines represent vector sequences. Probe A and Probe B were used on a northern blot having 2mg of poly A+ RNA (lane 1 and 2); and 10 and 20 mg of total RNA (lanes 3 and 4) prepared from wild type cells. The sizes of the transcripts were determined by probing the blot for known transcripts (STE2 and TCM1 ) and also by running DNA markers (1 Hindiil) on the gel.

c ^{X H}

pUV18-151"\..1'\/'i I

^{I I}

rvvv

probe A probe 8

poly A+ Total poly A+ Total

1 2 3 4 1 2 3 ⁴

2.3kb- . . . 2.3kb- . .

1.6kb- 1.8kb-

probe A probe 8

CHAPTER4

lntron stabilization in the yeast mutant prp26

ABSTRACT

We have recently reported the isolation of a novel splicing mutantprp26 whose phenotype is intron accumulation. Further analysis of this mutant phenotype reveals that the intracel- lular intron exists in a lariat and a circular form lacking nucleotides from the 3' end of the lariat. The stabilization of the cellular intron results from its association with a large particle having sedimentation properties similar to that of the spliceosome ( -40S). The composi- tion of this intron-containing particle was analyzed after gradient sedimentation by northern analysis for the presence of the intron and snRNAs. A co-sedimentation of spliceosomal snRNAs and the intron was observed. Antibodies that precipitate snRNPs: tri-methyl G cap and anti-Sm antibodies, immunoprecipitate the intron. Splicing extracts prepared from strains bearing this mutation are active for splicing but are defective in vitro in the degrada- tion of in vitro generated intron.

INTRODUCTION

The removal of intervening sequences from the primary transcript occurs by RNA splicing.

The study of pre-mRNA splicing in in vitro systems has facilitated the understanding of the mechanism of pre-mRNA splicing. A two step mechanism of splicing has been established (reviewed in Padgett et al., 1986; Green 1986). In the first step cleavage at the 5' splice junction occurs with concomitant formation of splicing intermediates- a lariat RNA of in-

tron-exon2 and exonl. The second reaction involves cleavage at the 3' splice site and liga- tion of the exons occurs. This cleavage ligation reaction releases the intron in a lariat form.

Although, mechanistically, the splicing scheme is similar to that of the self-splicing Groupll introns (Cech 1986), nuclear pre-mRNA splicing depends upon the assembly and function of an RNA and protein complex termed the spliceosome (Brody and Abelson 1985; Grabowski et al., 1985; Frendewey and Keller 1985; Perkins et al., 1986; Bindereif and Green 1986).

Splicing complexes were first detected upon sedimentation analysis of splicing re- actions as 40-60S particles. Small nuclear ribonucleoprotein particles (snRNPs) have been shown to be essential components for spliceosome formation and function (Steitz eta/., 1987; Guthrie and Patterson 1988). Purification of gradient isolated spliceosomes by affinity chromatography revealed the presence of U2, U4, US and U6 snRNAs (Grabowski and Sharp 1986). Identification of assembly intermediates of the spliceosome initially by sedimentation analysis and subsequently by non-denaturing gel electrophoresis has facilitated the development of an assembly pathway. An A TP-dependent binding of U2 snRNP is followed by the binding of U4, US and U6 snRNPs (Konarska and Sharp 1986;

Pikielny and Rosbash 1986). Prior to the first splicing reaction U4 leaves the complexes (Cheng and Abelson 1987; Lamond et al., 1988). Hence the splicing intermediates are found in a complex of U2, US and U6. Although the U1 snRNP is known to be essential for pre-rnRNA splicing, it was not detected on these assembly intermediates. Analysis of the kinetic order of binding of snRNPs to the pre-mRNA by affinity chromatography has implied an early and A TP independent binding of U1 to the pre-mRNA, which is followed by the binding of U2, U4, US and U6 snRNPs (Bindereif and Green 1987; Ruby and Abelson 1988).

The absence in vivo of branched RNA in the cytoplasm and low levels of branched RNAs detected in the nucleus imply a rapid degradation of the lariat intron (Wallace and Edmonds 1983). A specific enzymatic activity that hydrolyzes 2'-S' phosphodiester bonds has been characterized in HeLa cell extracts (Ruskin and Green 198S; Arenas and Hurwitz 1986). However, in vitro some of the lariat intron is associated with snRNPs in large complexes similar to the spliceosome. These complexes are thought to be involved in the dis-assembly of the spliceosome (Konarska and Sharp 1987; Lamond eta/., 1987). The release of the intron as a particle explains the protection of in vitro generated intron from the effective debranchase present in HeLa cell extracts. The mRNA is released from spliceo- somes and appears to be associated with hnRNP type complexes.

Analysis of splicing factors in yeast has the added advantage of the availability of temperature sensitive mutations (prp 2 -II ), which define protein components of the splic- ing machinery (reviewed in Warner 1987; Vijayraghavan and Abelson 1988). In vitro heat inactivation of extracts from these strains has implied a direct role for seven of the above complementation groups in the pre-mRNA splicing (Lustig et al., 1986). All but one of these complementation groups (prp2 ) defines components required early in spliceosome assembly (Lin et al., 1987). A recent search of more such mutants has obtained new com- plementation groups with novel phenotypes affecting late events in splicing. Two groups, prp26 and prp27, define gene products required after splicing is completed, as accumula- tion of only the excised intron is observed (Vijayraghavan et al., manuscript submitted). In this paper we have analyzed the biochemical effects of prp26, which causes accumulation of high levels of intron. We provide evidence for the stabilization of the intron in a large snRNP-containing complex, probably similar to the post splicing complexes observed in higher eukaryotic splicing systems.

RESULTS AND DISCUSSION

Accumulation of the branched intron

A novel yeast mutant with a phenotype of excised intron accumulation (prp26 ) was iso- lated upon screening for mutants in pre-mRNA splicing (Vijayraghavan et al., manuscript submitted). This phenotype results from a trans-acting mutation in a single recessive gene.

Primer extension analysis on the intron had indicated that the precision of 5' cleavage and branch site selection were unaffected. However, much less extension product was made from a primer hybridizing between the branch and the 5' splice site, than from the primer hybridizing downstream of the branch point. This suggested to us that some of the excised intron lacks nucleotides 3' to the branch site. In wild type cells no branched intron can be detected for at least four yeast transcripts tested. Only upon in vivo overproduction of an actin pre-mRNA fragment containing the intron, is the excised actin intron is detectable

(Domdey et

at.,

^1984). The excised intron consisted of four distinct forms- two of branched molecules and two linear forms of the intron.

To further characterize further the mutant phenotype in prp26 , we analyzed the in- tron after fractionation of total RNA from prp26 on 5% acrylamide, 8M urea gels to sepa- rate different forms of the intron. After electroblotting of the RNA, the membranes were probed with actin intron. They revealed the presence of three different forms of the actin intron in this mutant strain (Fig 1B). The largest form of the intron migrates at a position similar to that of the complete lariat form of the intron. The most abundant form of the in- tron has the same mobility as the circular form of the intron, which has lost some nu- cleotides 3' to the branch. A minor species of the intron, corresponding to a linear form, again lacking the 3' nucleotides, is also detected. These RNAs were assigned on the basis of their mobility and by comparison with published results (Domdey et al., 1984). None of these species is detectable in the wild type parental strain. That these forms of the intron are normal intermediates in the degradative process of the intron is indicated by the correla- tion between the forms of the intron seen in this mutant strain and those observed in vivo upon overproduction of actin intron containing fragment . The composition of the intra- cellular intron in prp26 is thus predominantly branched, although the lariat intron itself is not the major form of the intron. The presence of the circular intron probably results from nucleolytic digestion of the nucleotides 3' to the branch site. The small amount of the linear intron observed suggests that the normal degradative pathway functions to some extent in prp26.

Association of the intron in a 40S particle

The stabilization of the branched intron in the cell where it is normally degraded, suggests that it exists in an intracellular complex. One interpretation of the presence of the branched RNAs is that the enzymatic debranching function, responsible for degradation of the unique 2'-5' linkage in these branched molecules (Ruskin and Green 1985; Arenas and Hurwitz

1987), is defective in the mutant strain. Alternatively, protection of the released intron be- cause of a defect in the dis-assembly of the splicing components may be responsible for branched intron accumulation. We therefore investigated the possibility that this mutation prp26 results in a protection of the intron due to its association with spliceosomal compo-

nents.

Whole cell extracts prepared fromprp26 were fractionated on glycerol gradients. In parallel, deproteinized RNA from the same extract was also sedimented, to provide a con- trol for the migration of the naked RNA. Another gradient of an in vitro splicing reaction with labeled pre-mRNA provided a marker for the position of the 40S spliceosome. The gradient fractions obtained from the mutant extract or the naked RNA were individually ex- tracted and their RNA content analyzed on northern blots. These blots were probed with a labeled fragment of the yeast actin intron. A peak of the intracellular lariat intron, the cir- cular intron and the minor amount of the linear intron co-sedimented in the 40S region of the gradient (Fig. 1A ). In fact this intron containing particle sedimented nearly as fast as the spliceosome (data not shown). The presence of all forms of the intron in the same re- gion of the gradient suggested that they may exist in the same complex. In a parallel gradi- ent of the naked RNA the lariat and circular RNAs sedimented at the top of the gradient (Fig. 1B). These results show the intron is contained in a fast sedimenting particle of 38- 40S; the sedimentation coefficient suggests that the intron is retained in a large post-splicing particle.

Sedimentation of in vitro splicing reactions on gradients has demonstrated a co- sedimentation of the splicing intermediates in 40S spliceosome (Brody and Abelson 1985;

Grabowski et al., 1985; Frendewey and Keller 1985; Bindereif and Green 1986). The lar- iat intron is associated with both a large complex slightly lighter than the spliceosome (-40S) and in a heterogenous particle of about -20S (Brody and Abelson 1985; Bindereif and Green 1986). The mRNA is released as a smaller, more heterogeneous particle of about 20S. Analysis of the assembly of higher eukaroytic spliceosomes on non-denaturing

gels has also revealed the presence of a large intron- containing complex that most likely represents a post-splicing complex (Konarska and Sharp 1987). After gradient isolation the large intron-containing complexes have been reported to be insensitive to in vitro enzy- matic debranching (Bindereif and Green 1986).

The intron-containing complexes ( -40S) that accumulate in prp26 containing pre- dominantly branched RNAs suggests increased stability of a post splicing complex. The stabilization is therefore unlikely to result from a defective debranchase. The detection of mostly branched intron in the mutant strain prp26 , in large complexes, re-enforces the the- ory that the stabilization of the intron is due to association with spliceosomal snRNPs.

lntron-containing complex is associated with snRNPs Co-sedimentation of snRNAs and intron complex

The blots of the gradient fractions used to detect the intron-containing particles were probed for the spliceosomal snRNAs- U1, U2, U4, U5 and U6 (Fig. 2A, Band C). U6 and U5 snRNAs co-sedimentation with the peak of intron is evident. The sedimentation of U4 with the intron is at about 38S. The larger yeast snRNAs U1 and U2 also appear to co- sediment. However, the analysis of the association of this intron complex with snRNAs through co-sedimentation analysis may be artifactual; because co-sedimentation does not indicate that they exist in the same particle. In fact, this analysis is influenced by the fact that endogenous multi-snRNP complexes have been detected in wild type cells, which be- have as large complexes (Konarska and Sharp 1987; Cheng and Abelson 1987).

Immunoprecipitation of intron by anti-3 mG and anti-Sm antibodies

The association of snRNPs with the intron in prp26 was tested with snRNAJP specific an- tibodies in immunoprecipitation experiments. One of the antibodies used is directed against the 3mo cap of snRNAs and the other antibody tried was the human autoimmune serum, anti-Sm, which has been shown to cross react with, and thus precipitate, yeast snRNPs

(Riedel eta/., 1986; Siliciano et al., 1987). The anti 3mG or the control pre-immune serum bound to protein A-Sepharose beads were incubated with whole cell extracts prepared from the parent wild type strain or the mutant prp26 strain. The bound RNAs were released by phenol extraction separated on denaturing gels and blotted to nylon membranes. The intron RNA was immunoprecipitated from prp26 extract by both the anti_3mG and anti-Sm sera (Fig. 3). As expected, because wild type extracts do not contain introns, none are im- munoprecipitated. Antibodies directed against the PRP4 gene product of yeast specifically precipitate U4 and U6 snRNA from yeast extracts (J. Banroques and J. Abelson, manuscript in preparation). These antibodies were also used to investigate if PRP4 protein is associated with the intron particle, but no precipitation of the intron was observed (data not shown).

The other product of the splicing reactions, mRNA, is not associated with snRNAs and is thus not precipitated (Fig. 3) ( Bindereif and Green 1986; Konarska and Sharp 1987;

Lamond et al., 1987). The absence of the mRNA in the immunoprecipitates implies that the precipitation of the intron is probably not due to non-specific binding of RNA to the beads or the antibodies. The fact that the immunoprecipitation occurs by specific recogni- tion of snRNA cap epitope is indicated by the effective precipitation of the snRNAs U5 and U4 (Fig. 3) and the U1 and U2 snRNA (data not shown). The anti-Sm antibody used in these experiments largely recognizes the yeast U2, and to some extent the U1, snRNA, but does not precipitate the U5 and U4 snRNAs (this study; S.-C. Cheng and J. Abelson un- published data). These experiments together indicated a direct precipitation of cellular intron in prp26 , due to its association with snRNPs.

Immunoprecipitation of large intron-containing complexes from higher eukaryotic splicing reactions indicates the presence of Sm determinants in these complexes, while the smaller intron complexes are not precipitable (Bindereif and Green 1986). The association of U5, U6 and U2 with the excised intron of higher eukaroytic pre-mRNAs has been shown by gel electrophoresis (Konarska and Sharp 1987). Our finding that the endoge-